Circular No. 05/2026/TT-BKHCN on promulgation of the national artificial intelligence ethics framework

CIRCULAR

PROMULGATION OF THE NATIONAL ARTIFICIAL INTELLIGENCE ETHICS FRAMEWORK

Pursuant to the Law on Artificial Intelligence No. 134/2025/QH15 dated December 10, 2025;

Pursuant to Decree No. 55/2025/ND-CP dated March 02, 2025 of the Government on functions, responsibilities, jurisdiction and organizational structure of the Ministry of Science and Technology;

At the request of the head of the National Institute for Digital Technologies and Digital Transformation, and the head of Legal Department;

The Minister of Science and Technology shall promulgate the Circular on promulgation of the national artificial intelligence ethics framework.

Article 1. Scope and regulated entities

1. This Circular promulgates the national artificial intelligence ethics framework (hereinafter referred to as “ethics framework”).

2. This Circular applies to state management authorities, organizations and individuals that are developers, providers, operators or users of artificial intelligence (hereinafter referred to as “AI”) systems serving state management or public services.

3. Vietnamese and foreign organizations and individuals conducting research, developing, providing, deploying or using AI systems in Vietnam are encourage to apply this Circular.

Article 2. Definitions

For the purposes of this Circular, the terms below shall be construed as follows:

1. AI ethics is a system of values, principles and standards for research, development, provision, deployment and use of AI to ensure respect for human beings, human rights, public interests and sustainable development.

2. Data poisoning refers to attacks aimed at modifying or manipulating training datasets.

3. Model poisoning refers to attacks aimed at pre-training components of a model that is used during the training process.

Article 3. National AI ethics framework

1. Ensure the safety, reliability and harmlessness to human life, physical health, honor, dignity and mental health.

a) Safe by design: Organizations and individuals are responsible for anticipating scenarios that may cause harm to human life, physical health, honor and dignity, and for developing preventive measures.

b) Reliability and quality: Organizations and individuals are responsible for establishing quality criteria for data, models and outputs; developing mechanisms for testing, validation and internal verification before deployment.

c) Human control: Organizations and individuals are responsible for developing mechanisms for supervision and human intervention that are appropriate with the impact of systems; ensuring the maintenance of human control and intervention over all decisions and behaviors of AI systems.

d) Resilience, response and security: Organizations and individuals shall develop mechanisms for receipt of feedback, detection of faults and remediation; and have backup plans in cases where systems malfunction or are abused.

e) Security for AI systems: Organizations and individuals shall apply appropriate protection measures to prevent, detect, block and respond to intrusions, takeovers, data poisoning, model poisoning, adversarial attacks, vulnerability exploitation, data leaks and abuse of AI systems; ensure the confidentiality, integrity and availability of data, models, algorithms and relevant infrastructure.

2. Respect human rights and civil rights; ensure fairness, transparency and non-discrimination in development and use of AI.

a) Respect for human rights and civil rights: Organizations and individuals shall apply appropriate review measures to ensure that AI systems do not infringe upon privacy, personal data, freedom of will, access to information, equal treatment rights and other lawful rights as prescribed by laws.

b) Fairness and non-discrimination: Organizations and individuals shall apply measures to identify and mitigate data bias, model bias and operational bias; consider impacts on vulnerable groups including children, the elderly, persons with disabilities and disadvantaged groups.

c) Transparency: Organizations and individuals shall develop appropriate notifications regarding the use of AI; provide reasonable information about objectives, scope, data, general functioning methods and limitations of systems; and prevent any misunderstanding about capabilities of systems.

d) Explainability and accountability: Organizations and individuals shall identify potential impacts of systems, prepare explanatory documents and evidences regarding the design, training, and testing processes; and delineate the responsible parties accountable for decisions generated by systems.

3. Promote happiness, prosperity and sustainable development for individuals, communities and society.

a) Social benefit: Organizations and individuals shall identify public benefits, added values and positive impacts of systems on people and communities; and prepare plans to address and mitigate negative impacts before deployment.

b) Inclusive development: Organizations and individuals shall prioritize designing interfaces that are accessible and easy to use; and narrow the digital divide between regions and population groups.

c) Sustainable development: Organizations and individuals developing or deploying AI systems are responsible for considering energy consumption, computational resource use and environmental impacts throughout the lifecycle of systems; prioritizing technical solutions, infrastructure and operational procedures that conserve energy and reduce emissions.

d) Respect for culture and social values: Organizations and individuals shall design AI systems in accordance with social ethics and the cultural identity of Vietnam; and not create information that is discriminatory, prejudiced or detrimental to community interests.

4. Encourage innovation and social responsibility in research, development and application of artificial intelligence.

a) Innovation: Organizations and individuals shall conduct experiments, pilot projects and expansion of AI application with responsibilities; promote open research and knowledge sharing in accordance with laws on intellectual property protection.

b) Social responsibility: Organizations and individuals developing, deploying and using AI systems shall define responsibilities of subjects throughout the lifecycle of systems; ensure that there are hubs to receive, handle complaints and remediate consequences.

c) Capacity development and collaboration: Organizations and individuals shall focus on training cadres and employees in AI awareness, ethical risks and safe usage skills; strengthen cooperation and learn from international standards on AI ethics.

d) Cooperation and learning: Organizations and individuals shall strengthen international cooperation; participate in initiatives, standards and codes of conduct on AI ethics; leverage private initiatives to serve public interests in accordance with conditions of Vietnam.

Article 4. Effect

This Circular comes into force from March 10, 2026.

Article 5. Implementation

1. The national AI ethics framework shall be reviewed and updated every 03 years, or when there is a significant change to technology, laws or management practices.

2. The Ministry of Science and Technology shall take charge of the implementation of this Circular; provide guidelines for management and evaluation of the compliance; update the national AI ethics framework in accordance with the need for development and international standards; develop and operate digital platforms and automatic testing tools to assist ministries, central authorities, local authorities and enterprises in self-evaluating their compliance with AI ethics.

3. Authorities, organizations and individuals shall report any difficulties and issues arising during the implementation to the Ministry of Science and Technology for research and appropriate amendments./.